The Atlantic Treaty Organization

The Atlantic Treaty Organization is the defensive alliance of the Atlantic states, established by the Treaty of Brussels in 1949 and seated at Brussels since. It has seventeen members, has never fought a war, and has never once invoked the guarantee that is its reason for existing.

Its guarantee is unusually hard. An armed attack on any member places every member at war, automatically and without a decision to be taken at the time. There is no provision by which a member chooses its response, no vote to be held, and no discretion reserved to any government once the fact of an attack is established. The obligation was drafted in that form deliberately, and the constitutional argument it produced in 1949 was the fiercest of Truman’s ministry.

The alliance is not the League of Nations, which remains the global frame above it, and it is not a general association of the Crown’s realms. Membership rests on Atlantic geography and willingness, and on nothing else. There is no monarchy test and no constitutional test; republics accede on the same terms as kingdoms, and sharing a monarch with the United Kingdom of America confers no standing whatever. States outside the alliance are outside it either because they are not Atlantic, as with the Empire of Raj, the Empire of China, Japan and the Kingdom of Arabia, or by their own decision, as with the West Indies Federation, the Union of South Africa and Cuba.

The European Concert has no defence arm. Hard defence in Europe is the ATO’s, and the division has held since the Concert was founded.

Who accedes, and as what

The alliance admits states, and the question of what counts as a state has been answered twice in different ways.

The three monarchies of the Benelux Union are members individually. Belgium, the Netherlands and Luxembourg each signed the Treaty of Brussels in their own right, each ratified it in their own parliament, and each holds a seat and a veto in the Council. The Union is a confederation of sovereign monarchies whose competences in defence run to planning and no further; each member retains its own armed forces, and a confederation with no command has nothing to commit. That its political seat is also the alliance’s is a matter of the same city being convenient twice.

The Scandinavian Union acceded collectively, as one party, in the first enlargement. It could do so because the Treaty of Kalmar had given it an integrated defence command over the national forces of its five members, so there was a single military authority for the alliance to deal with. Sweden, Denmark, Norway, Finland and Iceland are therefore covered by the guarantee without being parties to the treaty, which is an arrangement no other member has and which the Council took two years to satisfy itself about.

The asymmetry is not a matter of size or standing. It follows from what each union actually holds. A confederation that plans is three members; a confederation that commands is one.

The founding

After 1945

The Second Great War ended in 1945 with a settlement concluded at Versailles, and it did not produce a peace so much as a rearrangement of who the danger was.

The defeated aggressor had been France, whose fascist regime had invaded Norway and Denmark and had been destroyed in the war it started. The Russian Republic had not been defeated, had not been an ally in any sense the western capitals recognised, and emerged from the war with the largest land army in the world, an authoritarian government built by Admiral Kolchak out of the wreckage of the Romanov empire, and a frontier that now ran along the eastern edge of Europe. Within three years it had the bomb.

The League of Nations had been reconstituted under the San Francisco Charter in 1945, with its Security Council seated at Kyoto, and was the global frame within which all of this was supposed to be managed. It was not adequate to the specific problem, and everybody involved understood why: a body containing both parties to a quarrel cannot be the instrument by which one of them is deterred. Russia held a permanent seat and a veto. Anything the League might have done about Russia, Russia could stop.

Truman’s problem

Harry S. Truman, the twenty-third Prime Minister of the United Kingdom of America, took office in 1945 at the head of a Labour and Social Democratic ministry, and the alliance is the principal achievement of his seven years.

The problem he set himself was narrow and he stated it repeatedly. A guarantee that committed nobody to anything would deter nobody, and the interwar years had produced a great many of them. A guarantee that created a permanent supranational military command, answerable to no parliament and capable of generating its own obligations, would deter Russia and would also place the war power beyond democratic reach in every country that signed it.

Truman regarded the second danger as the more likely and said so. His concern was not that generals would seize power but that a professional allied staff, doing its ordinary work of planning, would in time produce a set of arrangements that made war automatic in circumstances no government had chosen. He wanted the obligation to be absolute and the machinery to be incapable of triggering it.

Brussels, 1949

The treaty was negotiated at Brussels through the winter of 1948 and signed there on 4 April 1949.

Brussels was chosen because it was nobody’s capital in the sense that mattered. New Westminster would have made the alliance an American instrument in appearance as well as in fact; London and Berlin each carried too much history with the other members; and Belgium was a small monarchy with a serious diplomatic service, a recent occupation to remember, and no capacity to dominate anything. It has been the seat ever since.

Eight states signed: the United Kingdom of America, Great Britain, the German Empire, France, New Albion, and the three monarchies of the Benelux Union separately, Belgium, the Netherlands and Luxembourg.

France

France was the hard case, and the article states the difficulty plainly because the alliance never resolved it.

Four years before the signature at Brussels, France had been the aggressor power of the Second Great War. Its fascist government had invaded Norway and Denmark, occupied them, and been destroyed by the coalition that included every other founding member of the alliance. The Versailles settlement of 1945 had constituted Normandy as a semi-independent constituent duchy within the French Empire under a foreign Duke, the Hanoverian monarch, and placed the restored Bonapartist empire under conditions.

The case for including it was geographic and was unanswerable on its own terms. France is the ground between the Atlantic and Germany. An alliance for the defence of western Europe that excluded France had no depth, no ports on the Bay of Biscay, and no way of reinforcing the German frontier except across territory it did not control. The military members of every delegation said the same thing in different languages, which is that a line drawn on the Rhine is not a line that can be held.

The case against it was moral and was also unanswerable. The regime had changed; the country had not moved. The people who had been invaded were being asked to sign a mutual guarantee with the state that had invaded them, nine years after it had done so, and to promise to go to war automatically in its defence.

The result was that France was a founding member and the countries it had occupied were not. The Scandinavian states, which had concluded the Treaty of Kalmar in the same year, did not accede at Brussels and did not accede for some years afterwards, and the reason given in every Nordic account is France. The alliance’s founding document was signed by the recent aggressor and not by two of its victims, and no member has ever offered a defence of this beyond the one about depth.

What the treaty says

The text above is the treaty as signed in 1949. The protocol of accession of 1999, which admitted Nigeria, widened the treaty area: Article IV’s “in Europe, in North America” became “against the territory of any of the Parties, other than territories overseas”, and Article V’s definition of the area was rewritten to match. The guarantee therefore covers each member’s own territory wherever it lies, and Nigerian soil is guaranteed as fully as any other member’s, while the members’ overseas possessions remain outside it as the founders intended.

The two articles are read together and were drafted together. Article IV makes the obligation absolute and removes every subsequent decision from every government. Article VII ensures that the only thing capable of triggering it is an enemy.

This is the whole of Truman’s design. The alliance is slow because nothing it does can commit anybody; it is credible because once the enemy acts, nothing anybody does can uncommit them. The last sentence of Article VII was drafted by Truman personally, according to the American delegation’s papers, and replaced three longer versions proposed by lawyers.

The ratification

The treaty was ratified in eight parliaments through the summer of 1949 and came into force on 1 September. The argument in New Westminster was the fiercest of Truman’s ministry and turned on a single constitutional question.

Could Parliament bind its successors to a future war?

The Charter reserves the authorisation of war to Parliament. Article IV places the Realm at war automatically upon an event, with no vote to be taken at the time, and therefore appears to remove from every future Parliament the decision that the Charter reserves to it. The opposition case, argued by serious constitutional lawyers on both front benches, was that the treaty was in substance a permanent delegation of the war power to a foreign event, and that no Parliament could make it.

Truman’s answer was that Parliament was exercising the war power rather than surrendering it, and exercising it in the only form in which it could be made to deter anybody: once, in advance, deliberately, and in public. A guarantee that required a debate at the moment of attack was a guarantee an enemy could plan around. A guarantee given in advance was a decision Parliament had actually taken, on a question it had actually considered, with the consequences on the table.

The treaty was carried on that understanding, and the understanding was stated in the resolution. It has never been tested, because Article IV has never operated, and Irish, German and British constitutional writers have all observed that the argument is unresolved rather than settled. The Realm has committed itself to a war it has not yet had to fight, and nobody knows what its Parliament would do on the day.

Structure

The Council

The Atlantic Council is the alliance’s governing body. Every member holds one seat and one vote, and it decides by consensus, which means that every member holds a veto over everything the alliance does.

Everything, that is, except the one thing that matters. The guarantee is not a decision and cannot be vetoed. Article IV operates on an event, not on a resolution, and the Council has no power to declare that an attack has not occurred, to decline to act upon one, or to release any member from an obligation that arises without its participation. A member may veto a budget, an exercise, an accession, a communiqué, the appointment of a commander and the wording of a press statement. It may not veto a war that has already begun by operation of the treaty.

The consequence is an organisation that is easily paralysed in peacetime and cannot be paralysed at all in the only circumstance it exists for. Members have used the veto freely and often, over basing, over money, over enlargement and over whether particular exercises should take place at all, and none of that obstruction has ever touched the guarantee, because the guarantee was placed beyond the reach of the body that governs the alliance. Truman’s delegation regarded this as the single most important structural decision at Brussels and it is the one most often misunderstood by commentators who describe the alliance as consensus-bound.

The Council sits permanently at ambassadorial level and meets at ministerial level twice a year, and at head of government level when it chooses.

The Secretary-General

The Secretary-General is the alliance’s civilian head, chairs the Council, and holds the office for four years.

There is no convention about nationality. The post has been held by nationals of seven different members, including two Americans and three Germans, and the absence of a rule is deliberate: the founding negotiation considered a convention excluding Americans, on the ground that the alliance should not be visibly led by its largest member, and rejected it on the ground that a rule of that kind would advertise the anxiety it was designed to conceal.

The office is weaker than its equivalents elsewhere. The Secretary-General cannot propose a measure against the opposition of a member, cannot speak for the alliance beyond the terms of an agreed communiqué, and has no authority over any military formation. What the office actually does is manage the consensus: identifying which members can be brought to agree, in what order, and at what price, and preventing proposals from reaching the Council before they can carry. Successful holders have all been former foreign ministers of small members.

The Supreme Allied Commander

The Supreme Allied Commander is the alliance’s senior military officer, and the post rotates among the member nations.

This is the alliance’s most distinctive military arrangement and the one most criticised by professionals. A rotating command means that the officer responsible for allied planning changes nationality every four years, that each incoming commander brings a staff formed in a different service tradition, and that the continuity of allied planning depends on the permanent international staff rather than on the commander. Every serious study of the alliance’s military effectiveness has recommended a permanent commander. The Council has rejected the recommendation on each occasion.

The reason is Truman’s, and it has outlived him. A permanent supreme command would in practice be American, because the American contribution is the largest and no other member could sustain the appointment indefinitely; and an alliance whose forces are permanently commanded by the nationals of one member is that member’s army with allies attached. The rotation is inefficient by design. It is the price the alliance pays for not being an American command, and the members who complain loudest about the inefficiency are also the ones who would object most strenuously to the alternative.

In peacetime the Supreme Allied Commander commands nothing. The post directs planning, sets standards, runs exercises and maintains the assignment tables. No national formation is under allied command until Article IV has operated, and Article VII ensures that nothing the commander does can bring that about.

The forces

The alliance has no standing multinational formations. What it has is a system of earmarked national forces: units designated by their governments, held at agreed readiness, equipped to common standards, and exercised together, which remain under national command until the guarantee operates.

The system is the practical expression of Article VII. A German division earmarked to the alliance is a German division, commanded by German officers, paid by the German treasury and answerable to the Reichstag, and it becomes part of an allied formation only in a war that has already begun. Standardisation is the substitute for integration: common calibres, common fuels, common signals procedures, common map conventions and a common language of orders, so that formations that have never been commanded together can be commanded together at short notice.

Whether this works is unknown and unknowable. The alliance has exercised the transition to allied command hundreds of times and has never performed it. Its own staff studies say the transition would take between four and eleven days depending on the theatre, and the range in that estimate is the honest measure of what nobody knows.

The oldest quarrel

The alliance has argued about the sea and the land since 1949, and the argument has never been settled because both sides are right.

The maritime case is that the alliance is called Atlantic for a reason. Its European members cannot be reinforced except across the ocean, the American army that would reinforce them is three thousand miles away, and an enemy that cannot break the Rhine can win by sinking the convoys. On this reading the decisive theatre is the sea approaches, the decisive arm is anti-submarine warfare, and everything on land is a holding action until the crossing is secured.

The continental case is that the war would be decided long before any convoy arrived. The frontier that matters is the German eastern frontier and the Finnish border, an enemy attack would reach its objectives in days rather than months, and an alliance that invests in escort carriers instead of armoured divisions is preparing to reinforce a Europe that will already have fallen.

The argument maps almost exactly onto national interest, which is why it has never resolved. The Americans, the British and New Albion have generally argued the sea; the Germans, the Scandinavians and the Benelux monarchies have generally argued the land. The alliance has funded both, inadequately, for eighty years, and the standing joke among its planners is that the ATO has the second-best navy and the second-best army in the world and would need the best of one of them.

Nuclear weapons

There is no nuclear sharing in the alliance, and there never has been.

Two members hold nuclear weapons. The United Kingdom of America holds much the larger arsenal. The German Empire holds a modest one under a declared retaliatory doctrine, and holds it precisely because there is no sharing: a Germany that could not rely on American weapons being released for German defence built its own, and said so at the time. The alliance’s refusal to construct a sharing arrangement is therefore the direct cause of there being two nuclear powers in it rather than one.

The refusal followed from Article VII. A dual-key arrangement places a foreign hand on a national weapon and creates exactly the kind of standing allied machinery that Truman’s clause was written to prevent. It was proposed in the 1950s, examined for two years, and abandoned when the lawyers reported that no version of it could be reconciled with the treaty.

The Scandinavian Union is a full member and is deliberately nuclear-free, hosting no weapons and holding none, which is a position available to it only because no member is required to host anything.

Money

The German Empire and the United Kingdom of America carry the alliance in roughly equal measure, and this is the single most consequential fact about its internal politics.

An alliance funded overwhelmingly by one member is that member’s alliance, and the burden-sharing quarrel familiar from other arrangements, in which a dominant paymaster complains that its allies are free-riding, does not occur here in that form. What occurs instead is a rivalry: two large members, contributing comparably, each with a claim to shape the alliance’s priorities, and each supported by a bloc of smaller members whose interests align with one or the other. The sea-versus-land argument is in part the same argument conducted in strategic language.

Smaller members contribute proportionately and are not seriously accused of failing to. The alliance’s published cost-sharing formula is the least contested document it produces.

Languages, accession and partners

The working languages are English, French and German, and all three texts of the treaty are authentic.

A new member is admitted by the unanimous consent of the existing members and by ratification in each of their parliaments, which makes accession slow and gives every member a veto exercised twice. The requirement has been used to delay applicants and never, so far, to defeat one outright.

The alliance also maintains a formal partner status short of the guarantee, for Atlantic states that will not or cannot accede: consultation, intelligence exchange, exercise participation and standardisation, without Article IV. It is available on Atlantic geography alone. This is why the arrangements between the United Kingdom of America and Japan and the Kingdom of Arabia are bilateral treaties and not alliance partnerships: those states are not Atlantic, so no category within the alliance is open to them at any level.

Membership

Thirteen parties, seventeen states

The alliance has seventeen member states and thirteen parties to the treaty, and the difference is the Scandinavian Union, whose collective accession brought five nations under the guarantee through a single signature.

Component polities are covered through their sovereign and hold no separate standing. Brittany and Normandy are covered through France; England, Scotland, Wales and Kernow through Great Britain; the German kingdoms, grand duchies and free cities through the Empire. The European Concert applies a looser version of the rule, with an observer seat for Normandy and legal membership for the states inside its two confederations; the alliance keeps the strict form, because an alliance that admitted components would give some members a dozen votes and others one.

Normandy is the exception that proves it. The Versailles settlement of 1945 placed the Duchy under the French Empire’s reserved sovereignty while granting it a neutral exception from ordinary French commitments, so French membership does not automatically bind Norman formations. The alliance has never established what this means for Article IV, and the standing assumption at Brussels is that it means nothing, because an attack on Normandy would be an attack on French territory whatever Normandy’s formations were doing.

The enlargements

The alliance grew in four waves, each well separated, and it has admitted nobody since 1999.

The first enlargement brought the Scandinavian Union and Portugal. The Nordic accession was the more difficult and took two years to negotiate, partly over the mechanics of a confederation acceding as one party and partly for the reason described above: the Union contained two states that France had invaded, and the Nordic governments required arrangements that no other member had been asked to accept. What they obtained was a formal exchange of instruments with France, negotiated separately from the accession and never described as an apology by either side.

The second enlargement brought the Socialist Kingdom of Italy, whose membership is the alliance’s standing anomaly. A socialist monarchy joining an alliance directed against the great republic of the east was regarded at the time as either a diplomatic triumph or an absurdity, and the Italian government’s own explanation, that its quarrel was with Russian conduct rather than with republican government, has never satisfied anybody. Italy has been an awkward member ever since, has vetoed more Council business than any other state, and has never once been suspected of unreliability on Article IV.

The third enlargement brought Spain, after the change of government that made it possible, and it was the last European accession.

The fourth enlargement brought the Federal Kingdom of Nigeria in 1999, after the Gulf of Guinea piracy crisis, in which attacks on tankers off the Niger Delta closed the shipping lanes for months. Allied navies patrolled the gulf with the Nigerian navy, and Nigeria then asked to join. It is the alliance’s only African member, and its accession was decided on the ordinary test: the Gulf of Guinea is an arm of the Atlantic, and Nigeria was willing. The protocol of accession widened the treaty area to the territory of any Party other than territories overseas, because the treaty’s 1949 wording guaranteed only European and North American ground and would otherwise have left the new member’s own soil outside the guarantee it was joining. It is the most recent accession, and nothing has been added since.

Who is not in it

The Kingdom of Ireland is the awkward absence.

Ireland is constitutionally non-neutral, by the logic of the shared Crown rather than by choice: a realm sharing a monarch with belligerents cannot claim neutrality without dissolving the doctrine on which its own sovereignty rests. It is also Atlantic, willing in principle, and covered by no guarantee.

The reason is domestic. Accession would require an Irish government to sign a treaty committing Irish soldiers automatically to a war declared by nobody in Dublin, and no Irish government has ever been able to carry that. The obligation Article IV creates is precisely the obligation the Easter Settlement was written to prevent the Crown from imposing, and a Taoiseach proposing it would be asking the Oireachtas to do voluntarily what eighteen years of the dominion period were spent resisting. Successive governments have examined it and none has laid a bill.

The practical result is a state that would certainly fight, has no obligation to, and receives no guarantee. Irish defence planning assumes British and American cooperation that no instrument requires. Alliance planners assume Irish facilities that no treaty provides. Both assumptions are probably correct and neither is written down anywhere.

Four Atlantic states are outside by their own decision: the West Indies Federation, the Union of South Africa, the United Provinces of the River Plate and Cuba. Their non-alignment is a choice rather than a disqualification, and all four are eligible on the geographic test. Non-Atlantic states are outside for the plain reason that they are not Atlantic, which is why the Empire of Raj, the Empire of China, Japan and the Kingdom of Arabia have no standing in the alliance at any level, and why the arrangements between Japan, Arabia and the United Kingdom of America are bilateral treaties instead.

The alliance in the cold war

France withdraws from planning

France withdrew from integrated planning in the 1960s and did not return for a generation. It never left the treaty and never suspended Article IV, and its guarantee to the other members and theirs to it continued without interruption throughout.

The cause was Normandy. Allied planning treated Norman ports, airfields and coastline as part of the theatre available to the alliance, which they are, while the Versailles settlement denied France the sovereignty over them that every other member exercised over its own territory. The French position was that its allies could not simultaneously treat Normandy as French for planning purposes and as not French for constitutional ones, and that a country asked to defend ground it did not own was being asked to accept a settlement it had never agreed to.

The alliance’s answer was that the Versailles settlement was not the alliance’s to revise, which was true and unhelpful. France withdrew its officers from the allied staffs, declined to earmark formations, and continued to attend the Council, pay its assessment and honour the guarantee. The withdrawal was from the machinery and not from the obligation, which is a distinction Article VII makes possible: the planning is severable because the planning does not create the commitment.

France returned to integrated planning in the 1990s, quietly, without a settlement of the Norman question, which remains exactly where it was.

1968

The closest the alliance came to war was not a war and was not intended by anybody, and it is the episode that taught the members what they had signed.

In the autumn of 1968, during an internal crisis in the Russian Republic whose nature was not understood in the west for some years afterwards, Russian formations concentrated on the Finnish frontier in strength and without warning. The concentration was defensive in origin, a consequence of a struggle in Petrograd rather than a decision about Finland, and nobody in the alliance knew that at the time.

What the alliance discovered in the following seventy hours was the character of its own guarantee. Finland is covered through the Scandinavian Union’s collective accession. An armed attack on Finland would place all sixteen states at war with Russia automatically, immediately, and without any government taking a decision. There was no vote to be held, no Council resolution to be drafted, no national deliberation to be conducted, and no mechanism by which any member could have paused, qualified or declined.

The Council met continuously. It had nothing to decide. Several governments discovered, in the course of those three days, that their own war planning had assumed a decision point that did not exist, and at least two heads of government asked their law officers whether the treaty said what their foreign ministries were telling them it said. It did.

Nothing happened. The Russian formations dispersed over the following fortnight and the crisis was not publicly understood until the archives opened in the 1990s.

The consequences inside the alliance were considerable and are the reason the episode matters. The staff studies on the transition to allied command date from the reviews ordered afterwards. The permanent Council session dates from 1969. And the argument about whether an automatic guarantee is a deterrent or a trap, which had been a constitutional debate conducted by lawyers in 1949, became a practical question that every subsequent government has had to answer for itself.

The alliance’s official position, then and since, is that the guarantee worked exactly as designed: the Russians did not attack, and the reason they did not attack is that there was nothing to be gained by testing whether sixteen states meant it. The position is unfalsifiable and is probably correct.

May Day, 2006

The attack

On 1 May 2006 four coordinated bombings struck rail, air, water and civic assembly in New York. The May Day bombings were the worst attack ever suffered by a member of the alliance on its own territory.

Article IV was never invoked. It has still never been invoked. The reason is the most serious thing in this article.

What the treaty required

Article IV operates upon an armed attack and places every member at war with the attacking state.

The phrase is not decorative. The guarantee is directed at states, because in 1949 nobody drafting it imagined an attack by anyone else, and it produces a specific consequence: the treaty operates only when a state is identified as the attacker. An atrocity committed by people acting for nobody is not within Article IV at all, however many are killed.

So the question in May 2006 was not whether the alliance would act. It was whether the United Kingdom of America would say that a state had done it. If it did, the treaty would operate at once, with no vote to be taken anywhere, in seventeen countries, and nothing any government could do would stop it.

The days that followed

Intelligence on the origin of the explosive, and on money moving through legitimate banks, was read within days as implicating the Russian Republic. The material was genuine and the reading was honest.

Forces moved. An ultimatum was drafted.

The Atlantic Council sat continuously through that period, as it had in 1968, and as in 1968 it had nothing to decide. Seventeen governments waited to discover whether they were at war, on the basis of an assertion that only one of them could make, about evidence that only one of them held, in a decision none of them had any part in. Several allied governments asked to see the material. The answer given at the time was that the assessment was continuing.

Prime Minister Al Gore refused to act on it. The attribution collapsed within weeks, before any strike, and it was wrong. He did not tell the Realm what he had done. His authority broke in the Commons on a different disclosure, he survived a confidence vote narrowly, and he resigned on 11 June to spare his party a fight during a national emergency. That he had stopped a war became public only years afterwards, at the Investigation.

What this means for the alliance

The alliance’s proudest claim is that Article IV has never operated in eighty years. The claim is true. It is true because one head of government, of one member, declined to assert a fact that would have made it operate, on evidence that turned out to be mistaken, and lost his office for reasons connected to that period without ever explaining the part that mattered.

Every structural feature that the alliance regards as a strength was, in those weeks, a liability.

The guarantee is automatic, so no ally could have paused it. The Council governs by consensus, but consensus does not reach Article IV, so no ally could have vetoed it. Article VII ensures that only an attack can trigger the guarantee, which means the trigger is a question of fact rather than of judgement, and the fact was to be established by the government that had just been attacked, in the days immediately after it had been attacked, on intelligence it was not obliged to share.

The design assumed that identifying an attacker would be easy, because in 1949 an armed attack meant an army crossing a frontier and the identity of the attacker was not in doubt. It made no provision for an attack whose authorship was contested, uncertain, or wrong.

The alliance has done nothing about this. Article IV has not been amended, no procedure for allied verification of attribution has been created, and no member has formally proposed one. The reason usually given is that any verification procedure would introduce a decision point, a decision point is a delay, and a delay is precisely what Article VII was written to prevent an enemy from exploiting. The reason less often given is that the only member that could be asked to submit its attribution to allied scrutiny is the United Kingdom of America, and no European government wishes to open that question.

The sealed annexes of the Investigation, which contain the attribution and the intelligence behind it, are due to open in 2031. The campaign to open them early is led by historians and the press. Whether the alliance’s own conduct in those weeks is in them is not known.

The alliance at present

What it is for

The cold war has cooled, the balance of terror has slackened, and Russia is in flux: in October 2028 the Russian Republic freed its vote, restored its dynasty and became the Russian Empire. Nobody in the alliance has a settled answer to what it is now for, and the absence is its central problem.

Three answers are offered and none commands agreement.

The first is that the alliance is insurance against a Russia that may yet turn back, which most members privately believe and few state publicly, because stating it is an accusation against a government the same members are negotiating with.

The second is that the alliance has found new work: sea lanes, disaster relief, counter-piracy, and operations under League mandate. This is true and is not enough. None of it requires an automatic mutual guarantee, and an organisation that justifies a war obligation by pointing at flood relief has conceded the argument.

The third is that the alliance should be wound down, or converted into something looser, and it is argued seriously in several member parliaments. Its difficulty is that nobody can explain how a guarantee is withdrawn without the withdrawal itself being read as a message, and a message read in Petrograd.

Whether Russia could join

The question has been asked since the thaw and is discussed more often than the alliance admits.

The answer is that Russia is not an Atlantic state, and the membership test is Atlantic geography and willingness. It is a complete answer, it requires no government to say anything impolite, and everybody involved understands that this is why it is the answer that gets used. A rule that excludes on geography is a rule nobody has to defend.

The peace movement

Opposition to the alliance is old, was largest in the 1980s over nuclear weapons, and has never disappeared.

The movement of the 1980s was a mass phenomenon in Germany, the Netherlands and the Nordic countries, and it fixed the terms in which the alliance is still criticised: that a guarantee which cannot be stopped is not a guarantee but a mechanism, that deterrence purchased with the threat of annihilation is not peace, and that the citizens of sixteen countries have been committed to a war none of them will be asked about. It won nothing formally. It is generally credited with the Scandinavian Union’s decision to remain nuclear-free and with the abandonment of several basing proposals.

It is smaller now and is growing again, for the obvious reason. An alliance whose antagonist is dissolving is an alliance whose critics have their easiest argument in eighty years, and the peace movement of the 2020s asks a question the alliance’s supporters find genuinely difficult: if the threat has receded, what is the automatic guarantee still for, and against whom.

The work of peace

Standardisation

The alliance’s largest continuous activity, and the one with the most effect on the daily life of seventeen armed forces, is standardisation.

Because no formation comes under allied command until a war has already begun, everything that would ordinarily be achieved by serving together has to be achieved in advance by agreement. A German battery must be able to fire ammunition delivered by an American lorry driver who cannot read German. A Portuguese frigate must be able to take a signal from a Norwegian aircraft and understand it identically. A Belgian field hospital must be able to read a Spanish casualty tag.

The instruments that do this are the standardisation agreements, of which there are several thousand, ratified individually by each member and binding only on those that ratify them. They cover ammunition natures, fuel specifications, connector fittings, signals procedure, map projection and grid references, phonetic alphabet, casualty marking, prisoner handling, engineering tolerances for bridging, the format of an operation order, and the precise meaning of about four hundred words of English, French and German that mean slightly different things in each service.

Nothing about this is glamorous and all of it is the alliance. A member that has ratified the agreements and exercised against them can be plugged into an allied formation in days; a member that has not cannot be plugged in at all. The standardisation programme is what the alliance’s budget mostly buys, and its published measure of readiness is the ratification rate rather than any count of divisions.

The system has a known weakness, which its own inspectorate reports every year. Ratification is voluntary and unratified agreements are common, particularly among members whose domestic industries would bear the cost of conversion. The alliance’s answer has been to publish the ratification tables by member, on the theory that no defence ministry enjoys appearing at the bottom of a list its parliament can read. This works better than it should.

The exercise cycle

The alliance exercises continuously, on a cycle that has scarcely changed since the 1970s.

The largest is the annual transition exercise, in which the machinery of Article IV is rehearsed from the moment of a notional attack: the recognition of the fact, the transmission of the state of war to seventeen capitals, the passage of earmarked formations from national to allied command, and the assumption of authority by the Supreme Allied Commander. It is the only occasion on which the alliance practises the thing it exists to do, and its results are classified in every member.

What is not classified is the estimate the exercises produce. The transition to allied command is assessed at between four and eleven days depending on theatre, and this figure has improved by about a day and a half in fifty years, most of it in signals rather than in anything else. The alliance publishes the number. Its critics observe that an alliance whose enemy could reach its objectives in less time than its own command arrangements take to form has a problem it has never solved. Its planners answer, correctly, that the transition begins the moment an attack begins and not the moment it ends, and that the relevant question is whether formations can fight nationally while the transition runs. They can, and do, and that is what the standardisation is for.

Smaller exercises run throughout the year: naval escort work in the western approaches, air defence over the German and Nordic frontiers, mountain and cold-weather training in Norway, amphibious work in Portugal and Spain, and a rolling programme of staff exercises without troops.

Intelligence

Intelligence sharing inside the alliance is extensive, uneven and permanently contentious.

There is no allied intelligence service and there was never any prospect of one, for the same Article VII reason that prevented an allied command: a body producing allied assessments would in time produce allied obligations. What exists instead is an exchange, member to member, under agreements of varying generosity, coordinated through a small assessment staff at Brussels whose function is to circulate what members have agreed to circulate and to notice when they have not.

The exchange is asymmetric because the capabilities are. The United Kingdom of America collects more than the other fifteen members combined and shares a great deal of it. What it does not share is the material bearing on attribution, sources and methods, and this exclusion is written into the exchange agreements rather than practised informally.

The consequences of that exclusion were demonstrated in 2006 and are described above. The alliance has not revisited the arrangement since, and the assessment staff’s annual report has recommended a review in each of the last nineteen years.

The civilian side

The alliance maintains a permanent international staff of about two thousand at Brussels, drawn from the member states and paid from the common budget: the Secretary-General’s office, the assessment staff, the standardisation directorate, the exercise planners, the budget division, and a legal service whose opinions on the treaty are the nearest thing the alliance has to jurisprudence.

The legal service is more important than its size suggests. Because Article IV operates on a fact rather than a decision, and because that fact has never occurred, the entire law of the guarantee consists of opinions written by lawyers about a situation none of them has seen. Those opinions run to some thousands of pages, are circulated to members, and bind nobody. They are nonetheless what every government’s own law officers read first.

The alliance and the other institutions

The League of Nations

The alliance operates within the League of Nations framework and is not a rival to it, a point both bodies are careful to restate.

The formal position is that the ATO is a regional arrangement of the kind the San Francisco Charter contemplates, that it exists for collective self-defence, and that its guarantee is the exercise of a right the Charter preserves rather than a departure from it. The League’s Security Council sits at Kyoto, Russia holds a permanent seat and a veto, and the alliance’s founders were entirely candid in 1949 that this was the reason the alliance was necessary: a body in which the danger holds a veto cannot be the instrument for containing it.

Relations between the two are correct. The alliance has contributed formations to League mandates on several occasions, always through the voluntary contributions of individual members and never under Article IV, and the distinction is maintained pedantically because it matters: an allied deployment under League mandate is a coalition of the willing wearing an allied badge, and carries no obligation on anybody.

The relationship’s difficulty is structural and unresolved. If a member were attacked by another League member, the guarantee would place seventeen states at war with a state holding a Security Council seat, and the League would be presented with a war among its own members that its Council could not address because one party to it holds a veto. Every legal service in the alliance has written an opinion on this. No two of them agree.

The European Concert

The European Concert has no defence arm, and the division of labour between it and the alliance is the most successful institutional relationship in the Atlantic world.

The Concert, founded at Straßburg in 2027, handles the political and economic integration of Europe. The alliance handles hard defence. Neither trespasses. The arrangement works because it was designed by people who had watched the alternative: a Europe with two overlapping security structures, in which every question could be taken to whichever body was likelier to give the answer a government wanted.

The relationship is nonetheless a live argument in several members, because the Concert is European and the alliance is Atlantic, and the difference is the United Kingdom of America. A European defence capability inside the Concert would be a capability without American participation, which is precisely why some Europeans want one and precisely why the alliance’s supporters oppose it. The Concert’s founding settlement excluded a defence competence deliberately. Whether it will stay excluded is one of the open questions of the 2020s, and the alliance’s own view of the matter is that a European pillar inside the alliance is welcome and a European alternative to it is not.

The bilateral treaties

The United Kingdom of America maintains bilateral defence treaties outside the alliance, of which the most substantial are with Japan and the Kingdom of Arabia.

These exist because those states are not Atlantic and no alliance category is open to them at any level, not because they are unsuitable or unwilling. The treaties provide consultation, intelligence cooperation, training, access and extended deterrence, and they carry no automatic obligation of the Article IV kind in either direction. They are American commitments and not allied ones, and no other member of the alliance is bound by them.

This produces a structural oddity that the alliance’s smaller members raise periodically. The largest member has substantial commitments in a region where none of its allies has any, and a war arising from those commitments would engage American forces earmarked to the alliance without engaging the alliance at all. The American answer has been that Article IV is a defensive guarantee and cannot be triggered by an American war of choice anywhere, which is correct as a matter of law and does not address the question of what would be left in Europe.

Bases, forces and manpower

The basing settlement

The alliance owns nothing. Every installation used by allied forces is national property, held by the state it stands in, and made available under a bilateral status agreement between that state and the state whose forces use it.

This follows from Article VII as everything else does. An alliance holding territory in its own right would be a standing entity with interests of its own, and the founders would not create one. What exists instead is a network of national bases with allied functions: American formations in Germany and the Benelux monarchies, British and American maritime facilities in Portugal and Iceland, allied air defence installations along the German and Nordic frontiers, and a system of pre-positioned stores that belong to the state that paid for them.

The status agreements are the alliance’s most persistent source of ordinary friction. They govern jurisdiction over allied personnel accused of crimes, liability for damage, customs and tax treatment, environmental standards and the rights of local employees, and every one of them has been renegotiated repeatedly under domestic political pressure. Nothing in the alliance generates more constituency correspondence, and nothing is less discussed at the strategic level.

The largest concentration is in Germany, for the obvious geographic reason, and the German arrangement is the most politically sensitive. Germany hosts more allied personnel than any other member, has done since 1949, and the question of what those forces are for once the antagonist has dissolved is asked more sharply in the Reichstag than anywhere else.

Manpower

The members do not maintain their forces on a common model and the alliance has never attempted to impose one.

Conscription persists in some members and has been abandoned in others, and the pattern does not follow either geography or politics. The Scandinavian Union conscripts and regards it as a civic institution rather than a military necessity. Germany ended peacetime conscription in the postwar decades and keeps a professional volunteer force. The United Kingdom of America has maintained an all-volunteer force since conscription was abandoned after Indochina in the 1970s, and regards it as constitutionally objectionable in peacetime. Portugal and Spain abandoned it within a decade of each other. Italy retains it in a form that its own general staff describes as ceremonial.

The alliance’s planners have complained about this for eighty years, because a force generation model that differs in every member makes mobilisation timelines impossible to standardise. The Council has never touched it, and would not be permitted to. Whether a state raises its army by ballot, by contract or by obligation is exactly the sort of question that Article VII places beyond allied reach, and the members that complain most about the resulting untidiness would be the first to veto a common standard.

What the members actually field

The alliance’s combined order of battle is large, unevenly distributed and misleading in aggregate.

The United Kingdom of America provides much the largest single contribution, in every arm, and provides almost all of the strategic lift and the great majority of the maritime capability in the western Atlantic. The German Empire provides the largest European land contribution and the bulk of the continental air defence. Great Britain provides maritime capability, particularly anti-submarine, and the alliance’s second strategic lift capacity. The Scandinavian Union provides forces optimised for a single theatre it knows better than anybody, and would be difficult to employ anywhere else. The smaller members provide contributions that are individually modest and collectively substantial, and in several cases specialised: Portuguese and Spanish maritime patrol, Belgian and Dutch engineering, Luxembourg’s disproportionate contribution to signals and to the common budget.

Adding these together produces a figure larger than any other military combination on earth, and the figure means less than it appears to. The forces are national, they are not commanded together, and the aggregate exists only in the event that the guarantee has already operated.

The alliance in domestic politics

In the United Kingdom of America

The alliance is not a live domestic controversy in the Realm and has not been one for decades.

It is understood as Truman’s achievement, is defended by all the main parties, and is criticised chiefly on cost and on the distribution of the burden. The constitutional argument of 1949, about whether Parliament could commit its successors to a future war, is taught in law faculties and does not appear in political debate, largely because the question has never been forced. The Alternative for the Realm and elements of the left have both argued at various times for a reduced commitment, and neither has made it a central position.

What has changed since 2006 is subtler. The May Day period established, in the minds of a generation of American officials, that the guarantee is a constraint on the Realm as well as a protection for it: that an American government which asserts a fact starts seventeen wars, and that this is a reason for extreme caution about assertions. Nobody has said this in public. It is visible in the drafting of American statements on attribution, which have been noticeably more careful since 2006 than before.

In Germany

Germany is where the alliance is most argued about, and the argument is old.

The German position has always been that the country hosts the alliance’s largest foreign presence, contributes on a scale comparable with the United Kingdom of America, and would be the theatre of any war the alliance fought. From that follows a permanent German claim to influence that the alliance’s structure does not straightforwardly provide, since every member has one vote and the Supreme Allied Commander rotates.

The German nuclear deterrent, built because the alliance would not share, is the sharpest expression of this. So is the recurring German argument for a permanent Supreme Allied Commander, which Germany has proposed four times and which every other member has understood, correctly, as a proposal that the post should be German rather more often than one time in seventeen.

The peace movement of the 1980s was largest in Germany and its residue is still present in German politics. The Alliance 90 and Green Realm tradition in its German form has been the most consistent critic of the automatic guarantee anywhere in the alliance, and its argument, that a country cannot be democratic in respect of the most serious decision a state makes if the decision was taken in 1949, has never been satisfactorily answered by anybody.

In the Nordic countries

The Scandinavian Union’s membership is settled, popular and conditional in a way no other member’s is.

The conditions are the nuclear-free position, which is absolute and not negotiable; the arrangement by which the Union accedes collectively, which preserves the Nordic Council’s authority over its own forces; and an understanding, never written down, that the alliance does not plan for the use of Nordic territory in ways the Union has not agreed. All three are respected. All three would be difficult to enforce if they were not.

Finland’s position is the most exposed in the alliance, since the Finnish border is the frontier the Union was built to hold and the place where the guarantee would most plausibly operate. Finnish opinion on the alliance is the most favourable of any member, which is not a coincidence.

In France

French opinion of the alliance is complicated by everything in French history since 1940, and the complication is not discussed candidly in French politics.

France is a founding member, was the aggressor power of the war that preceded the founding, and joined an alliance most of whose other founders had fought against it. The republican and imperial traditions in French politics have both preferred to treat the alliance as a matter of national interest rather than of national history, and the Norman question, which was the cause of the withdrawal from integrated planning, is the one place where the older argument surfaces.

The French position on European defence is accordingly the most enthusiastic in the alliance and the most suspect to the other members. A European capability inside the Concert would be a capability in which France’s weight was greater and its history less relevant.

In Ireland

Ireland is not a member, and the alliance is nonetheless a recurring subject in Irish politics, because Irish non-membership requires more explanation at home than membership would.

Ireland is not neutral, cannot be neutral, sits astride the western approaches, and would be a theatre in any Atlantic war. Its facilities are assumed by allied planners and its cooperation is assumed by everybody. What it does not have is a guarantee, a seat, a vote, or any say whatever in the arrangements that would govern a war fought around it. Irish commentators describe the position as the worst of the available options, and Irish governments continue to prefer it, because the alternative requires an Irish signature on an automatic commitment and no Irish government has been able to contemplate that since 1949.

The offices and their holders

The Secretaries-General

Fourteen people have held the office since 1949, for terms of four years, and the pattern of appointments is more revealing than any of the individuals.

The office has gone to nationals of seven members. Two have been American and three German, which is fewer than the size of those contributions would suggest and more than the smaller members would prefer. Nine of the fourteen were former foreign ministers, and eleven came from members other than the two largest. The consistent finding of every study of the office is that its successful holders have come from middle-sized members with no stake in the sea-versus-land quarrel, because a Secretary-General identified with either side of it cannot broker the other.

Three holders are remembered for particular reasons. The second, a Belgian, established the practice of never bringing a proposal to the Council until it could carry, which is the office’s central technique and has been followed by everybody since. The seventh, a Norwegian, chaired the Council through the crisis of 1968 and the reviews that followed, and the permanent session and the transition studies are both attributed to her. The eleventh, a German, presided over the beginning of the thaw and made the only serious attempt in the alliance’s history to define what it would be for afterwards; the paper he circulated was received politely by every member and adopted by none.

The office has no power to compel and its holders describe it in almost identical terms in their memoirs. The Secretary-General’s authority is the authority of the person who knows what seventeen governments will accept, and it lasts exactly as long as that knowledge is accurate.

The Supreme Allied Commanders

The military command rotates on a four-year cycle, and twenty officers have held it.

The rotation has not been strictly even and was never intended to be. Members with negligible forces do not supply supreme commanders, and the post has in practice circulated among six or seven of the larger contributors, with American and German officers holding it more often than others. The Council appoints by consensus and the appointment is one of the occasions on which the veto is used seriously.

The office is professionally uncomfortable. A supreme commander arrives with four years, a staff of seventeen nationalities, no forces, no authority over anything until a war begins, and a body of plans inherited from a predecessor who thought about the problem in a different service tradition. The permanent international staff provides the continuity, and the standing joke among them is that the commander’s principal function is to be the person whose signature the plans require.

The most consequential holder was the fourth, a British admiral, whose insistence that the transition to allied command be exercised rather than assumed produced the first transition exercise in 1961 and, seven years later, the only reason anybody had any idea what would have happened in the crisis on the Finnish frontier.

Symbols

The emblem

The alliance’s emblem is a white four-pointed compass rose within an unbroken white circle, on a field of Atlantic blue, with four narrow rays running from the star’s points to the ring.

The circle is the guarantee. A ring has no gap, no beginning and no end, and the specific thing it cannot do is be opened, which is what Article IV is. This reading was attached to the design at the time of adoption rather than afterwards, and appears in the Council minutes of 1949.

The compass rose is the argument.

A compass is a maritime instrument. The 1949 design competition produced fifty-one submissions, and the Council rejected a sword as too martial, an eagle as German or American depending on its treatment, a star as American, and an anchor as too narrowly naval. What carried was the compass rose, proposed by the British delegation, and it carried over the recorded objection of the German, Belgian and Dutch delegations, who observed that the alliance had not yet decided whether it was a maritime alliance or a continental one and that its badge had just decided for it.

They were right, and the objection has never been withdrawn. The sea-versus-land quarrel described above was already beginning in 1949 and has never been settled, and the emblem is the one place where the maritime case is not merely argued but assumed. The continental members have raised the design formally three times, most recently in the 1980s, and have been defeated on each occasion by the argument that changing an alliance’s emblem is a statement about the alliance and not about its badge.

The German army’s field publications refer to it, unofficially and universally, as the sailors’ star.

The blue is specified as Atlantic blue and is defined in the alliance’s standardisation agreements to a tolerance, which is the only aesthetic question the standardisation directorate has ever been asked to settle. It was settled after eleven months.

The wordmark, and the order of the languages

The emblem appears in formal use beside the alliance’s name in all three working languages, set at the same size, in the same weight, with none given prominence over the others.

ATLANTIC TREATY ORGANIZATION ORGANISATION DU TRAITE DE L’ATLANTIQUE ORGANISATION DES ATLANTIKVERTRAGS

The French line carries no acute on the second word. This is not an error and is not a concession to the engraver: the alliance’s style rules follow the ordinary French practice of omitting accents from capitals, which was near-universal in institutional lettering of the period and remains the rule for the alliance’s own inscriptions, plaques and letterheads. Set in lower case, in the body of a document, the word takes its accent like any other.

The equality of treatment is required and is checked. The alliance’s own style rules specify that no language may be set larger, first by weight, or above the others by any device of spacing, and the international staff has corrected member governments and contractors on the point for eighty years.

What could not be made equal was the order. Three lines have a first line, and in 1949 there was no principle available that did not favour somebody. Alphabetical order gives a different sequence in each of the three languages, which was pointed out by each delegation in turn as soon as it discovered that the ordering proposed was alphabetical in a language other than its own. Order of accession was identical for all three. Size of contribution was the argument the alliance was trying to avoid.

The sequence was drawn by lot at Brussels in 1949 and has never been changed. The draw was conducted by the Belgian foreign minister in the presence of the eight delegations, the result was recorded in the minutes without comment, and the order it produced is the order still used.

This is the alliance in miniature, and its own officials say so. A question with no principled answer was settled by a method nobody could object to, the result was written down, and everybody has abided by it for eighty years because the alternative to an arbitrary rule that works is an argument that does not end.

The name

The alliance is the Atlantic Treaty Organization, and the short form ATO is used everywhere including in its own documents.

The name was chosen for what it excludes. It carries no reference to defence, to the west, to freedom, to monarchy or to any political principle, because every such formulation was objectionable to somebody at Brussels and because the founders wanted a name that made no claim beyond geography. Membership is Atlantic geography and willingness, and the name says so and says nothing else.

The formulation used for the cold-war side, as distinct from the alliance itself, is the Allied monarchies or simply the Allies, and the two terms are not interchangeable. The alliance contains republics and always has. Describing the ATO as a monarchical bloc is an error that Irish, German and Portuguese officials correct with some sharpness, because it implies a membership test that does not exist and never has.

The crises

The internal quarrels

Beyond the crisis of 1968 and the events of 2006, the alliance’s serious internal disputes have been four.

The command question, which is permanent. Germany has proposed a permanent Supreme Allied Commander four times, most recently in the 1990s, and has been refused four times. The refusals have been argued on Truman’s ground and decided on the arithmetic that a permanent commander would be German or American and that neither outcome commands a consensus.

The French withdrawal from integrated planning, described above, which lasted a generation and was resolved by France returning rather than by anybody conceding.

The nuclear sharing proposal of the 1950s, examined for two years and abandoned when the legal service reported that no version of it was compatible with Article VII. The German deterrent followed within a decade, and the alliance has never formally acknowledged the connection.

The basing renegotiations of the 1980s, which coincided with the peace movement and produced the withdrawal of several proposals under domestic pressure in Germany, the Netherlands and the Nordic countries. It is the only occasion on which public opposition has changed an alliance decision, and both sides have reasons for not describing it that way.

What has never happened

No member has ever left the alliance. No member has ever been expelled, and there is no provision for expulsion. No member has ever been formally accused by another of failing an obligation under Article IV, because Article IV has never operated and there has been no obligation to fail.

The treaty contains a withdrawal clause requiring one year’s notice, which has never been used. Its existence is mentioned in every domestic debate about the alliance and its non-use is the alliance’s most quoted statistic after the number of times the guarantee has been invoked.

Industry and procurement

The armaments question

The alliance does not buy anything. It specifies, and its members buy.

The standardisation agreements determine what allied equipment must be capable of, and each member then procures from its own industry or from another member’s. The alliance has no common procurement agency, no common budget for equipment beyond its own infrastructure, and no power to direct an order anywhere. This follows from Article VII exactly as the absence of a common command does, and it has the same consequence: a function that would ordinarily be performed by an institution is instead performed by negotiation among seventeen governments.

The results are mixed and are honestly reported by the alliance’s own inspectorate. Standardisation of consumables, ammunition, fuel and connectors has been extremely successful, because the cost of conversion is low and the benefit is obvious. Standardisation of platforms has been a failure. The alliance fields, at present, six main battle tank designs, nine combat aircraft types and eleven classes of escort vessel, and every attempt to reduce those numbers has foundered on the fact that reducing them means somebody’s factory closing and somebody’s parliament being asked to vote for it.

The two industrial powers

The American and German defence industries are the largest in the alliance and are direct competitors, which shapes a great deal that is not usually described in industrial terms.

The sea-versus-land argument has an industrial dimension. Maritime capability is disproportionately American and British; heavy land capability is disproportionately German. A strategic decision in favour of either is also a procurement decision worth a great deal of money, and both governments are aware of this while neither states it. Alliance planners regard the industrial overlay as the chief reason the strategic argument has never been settled on its merits: there are no merits available that do not also decide a commercial question.

The smaller members have made a virtue of specialisation, partly by choice and partly because they cannot compete across the range. Belgian and Dutch engineering, Portuguese and Spanish maritime patrol, Luxembourg’s signals work and Nordic cold-weather equipment are each dominant in their niches across the whole alliance, and the smaller members’ industrial position is consequently more secure than their size suggests.

Assessment

The case for the alliance

The case is short and is difficult to answer.

The alliance was formed in 1949 against a specific danger: a Russian state with the largest army in the world, an authoritarian government, a nuclear weapon within three years, and a frontier running down the eastern edge of a Europe that had just been devastated. In the seventy-nine years since, that danger has not materialised. No member has been attacked by Russia. No war has been fought in Europe between great powers. The alliance’s guarantee has never been invoked because the circumstance it exists for has never arisen.

Its supporters say this is cause and effect, and the burden of proof is genuinely on the other side. An automatic guarantee removes from an aggressor every calculation on which limited aggression depends. There is no member to be isolated, no coalition to be split, no government to be deterred from responding, and no gap between an attack and a war. What Russia would have done facing sixteen states that had to be fought all at once, none of whom could choose not to fight, is not knowable, and the fact that it did nothing is the only evidence available.

The case against

The case against is that the alliance has never been tested and that everything claimed for it is therefore unproven.

Its critics observe that the guarantee has not operated, the transition to allied command has not been performed, the forces have never been commanded together, and the one attack a member actually suffered fell outside the treaty entirely. Eighty years of preparation have produced an organisation whose central mechanism is a legal opinion. Whether it works is an article of faith held with great confidence by people who have no way of knowing.

They observe further that the alliance has cost seventeen countries an enormous amount of money, has placed foreign forces in several of them for three generations, has committed the citizens of all of them to a war none of them was asked about, and has justified all of this by reference to an attack that never came, which is the one form of justification that can never be falsified.

The failings

Three should be stated plainly.

The alliance was founded with the recent aggressor inside it and two of its victims outside. France signed at Brussels in 1949 having invaded Norway and Denmark nine years earlier, and the Scandinavian states acceded later, and the reason given in every Nordic account is France. The alliance has never offered a defence of this beyond the argument about strategic depth, which is a real argument and is not an answer to the objection.

The guarantee was useless in the only attack a member has ever suffered. Article IV requires an attacking state. The May Day bombings killed more people on a member’s territory than any event since the war, and the treaty had nothing to say about them, because the drafters of 1949 could not imagine an attack whose author was not a government. This is a failure of foresight and would be forgivable. What is not forgivable is the second part.

And the alliance has done nothing about it since. Twenty-two years after May Day, Article IV has not been amended, no procedure exists by which allies might examine an attribution before being committed to a war by it, no member has formally proposed one, and the assessment staff’s recommendation of a review has been repeated and ignored in nineteen consecutive annual reports. The reason is not inertia. It is that any verification procedure would require the largest member to submit its intelligence to allied scrutiny, and no European government has been willing to ask, because asking would receive an answer.

The consequence is that the safety catch on an automatic seventeen-state war remains exactly where it was in 2006: in the judgement of whichever government has just been attacked, at the moment when its judgement is least reliable. In 2006 that judgement was exercised well, by a Prime Minister who lost his office for reasons connected with the period and never publicly explained the decision that mattered. The alliance’s institutional response to having been saved by one man’s restraint has been to continue relying on it.

The verdict

The alliance has held for seventy-nine years, has presided over the longest period without a great-power war in Europe in the modern era, and cannot demonstrate that it caused any of it.

It works in the sense that the thing it was built to prevent has not happened, which is the only sense available and is not nothing. It has never been required to work in the sense its members would recognise on the day, and the honest position, held privately by a great many of its own planners, is that nobody knows whether it would.

What can be said with confidence is narrower. The alliance made a specific bargain in 1949: absolute obligation, minimal machinery, no supranational command, and a trigger that no plan and no staff could pull. That bargain has been kept. It has produced an organisation that is slow, quarrelsome, easily obstructed and structurally incapable of being led by its largest member, and those are the features its founders intended rather than the defects its critics suppose. Whether an alliance built on those terms would fight is unknown. That it has never had to is either its vindication or the reason nobody can claim one.

If the guarantee operated

The first hours

The alliance has rehearsed this several hundred times and performed it never, and the following is what its own exercises assume.

An armed attack occurs. Article IV operates at that moment, by its own terms, and every member is at war from the instant the attack begins rather than from any subsequent recognition of it. There is no message that starts the war and no authority that declares it. What follows is not a decision but an administrative fact working its way through seventeen governments.

The attacked member notifies Brussels. The Council convenes, not to decide anything but to record. The Secretary-General transmits the fact to the capitals. Each government then does two things: it informs its own legislature that the state is at war, in the form its own constitution requires, and it releases its earmarked formations to the Supreme Allied Commander.

The releases are the part nobody is sure about. Each is a national act, performed by a national government under its own law, and the alliance has no power to compel any of them. A member is at war whether or not it releases anything. Whether it fights is a separate question that the treaty does not address, and the honest description of the guarantee is that it makes war automatic and participation national.

The four to eleven days

The transition to allied command is assessed at four to eleven days depending on theatre, and during that period the alliance does not exist as a military organisation.

Formations fight nationally, under national command, to national plans, against an enemy that has chosen the moment. The standardisation programme exists so that they can do this in mutually intelligible ways, and the alliance’s planners regard the first week as the period in which standardisation is worth more than command. Whether it is worth enough is the question the exercises cannot answer, because an exercise has a start time that everybody knows in advance and a war does not.

The legislatures

Every member’s parliament will be informed that its country is at war, and none of them will be asked.

This is the constitutional consequence of the 1949 bargain and it applies in seventeen countries with seventeen different constitutional traditions. In the United Kingdom of America, Parliament’s role was exercised at ratification and the Charter’s reservation of the war power is understood to have been satisfied in advance. In Germany, deployment abroad requires a prior resolution of the Reichstag, and the relationship between that requirement and an automatic state of war has never been resolved. In the Nordic countries the position is complicated further by the Union’s integrated command sitting between the national parliaments and the alliance.

Every legal service in the alliance has written on this. None of the opinions has been tested and none binds anybody. It is the single largest body of unexamined law in the Atlantic world, and it concerns what happens in the first week of a war among great powers.

The unanswered questions

The alliance has accumulated a set of questions it has never resolved, several of which it has deliberately declined to resolve. They are collected here because Irish, German and American constitutional writing treats them as a single subject.

Who establishes that an attack has occurred? Nobody. The fact establishes itself, and the alliance has no organ competent to find it, dispute it or review it.

What if a member says an attack has occurred and the others disagree? Unresolved. The treaty admits of no disagreement, because it creates no procedure within which disagreement could be expressed.

What if the attribution is wrong? Demonstrated in 2006 and unaddressed since.

What if the attacker is not a state? Outside Article IV entirely. Demonstrated in 2006.

What if a member refuses to release its forces? It remains at war and contributes nothing, and the alliance has no remedy. Every planner assumes this would not happen and none can say why not.

What if the attacked member does not want the guarantee to operate? The 2006 case in a different form, and the only one for which there is a practical answer: the member declines to assert the fact, and the treaty is silent because it has not been engaged.

What if two members go to war with each other? The treaty does not contemplate it. The legal service’s opinion is that Article IV would operate in both directions simultaneously, producing an incoherence it describes, in the only joke in forty years of alliance jurisprudence, as beyond the competence of this office.

What happens to Normandy? Unresolved since 1949.

Could the alliance be wound up? Yes, by the withdrawal of its members under the one-year clause, or by agreement. Neither has been attempted, and the difficulty is not legal. A guarantee that is withdrawn is a message about the conditions under which its members will no longer fight, and there is no way to send that message to only the intended recipients.

The view from Petrograd

What Russia says about it

The Russian Republic’s account of the alliance is consistent, has changed very little since 1949, and is not merely propaganda, which is what makes it difficult.

The Russian position is that the ATO is an encirclement. Russia emerged from the Second Great War having borne enormous losses, found itself facing a coalition of the industrial powers of the west formed four years later, and observed that the coalition included the country that had invaded Norway and Denmark and excluded the country that had done most of the dying. It notes that the alliance was formed before Russia had done anything to any of its members, that it has never been dissolved despite the disappearance of the circumstances that produced it, and that its guarantee is automatic in a way no other alliance in history has been.

The strongest Russian argument is the one western writers find hardest, and it is about the automatic guarantee itself. An alliance whose members cannot choose not to fight is an alliance that has removed the possibility of a limited quarrel. Any incident on the Finnish or German frontier, however local, however accidental, however swiftly regretted, becomes a war among great powers by operation of a treaty signed before most of the people involved were born. Russia has argued since 1949 that this is not deterrence but a hair trigger, and that the west has mistaken the removal of its own discretion for the removal of the enemy’s incentive.

The alliance’s answer is that this is exactly the point, and that an aggressor who cannot arrange a limited quarrel will not attempt one. Both propositions were true simultaneously for seventy-nine years and nobody can say which was doing the work.

Russia has no equivalent

Russia never built an alliance. It built a system of clients, bilateral treaties, basing agreements, subsidised armies and dependent governments, and at no point did it construct anything resembling a mutual guarantee among equals.

The reason is structural and is the most revealing thing about the Kolchak state and its successors. A mutual guarantee requires each party to accept an obligation to the others that it cannot subsequently escape, which requires each party to be regarded as an equal whose survival is worth the others’ war. Russia has never regarded any of its partners in that way. Its arrangements run outward from Petrograd and not between the states at its periphery, and a client that was attacked received whatever Russia decided to send, which is a relationship of protection rather than of alliance.

The consequence is that the cold war was not symmetrical, and the standard western description of two blocs facing each other is inaccurate. On one side sixteen states, all of which had to fight if any was attacked, none of which could be picked off. On the other, one great power and a periphery of dependants, any of which could be abandoned at Petrograd’s discretion and several of which were.

Western writers have made a great deal of this and Russian writers do not dispute the facts, only the inference. The Russian answer is that an empire does not need an alliance, which is candid, and that the ATO’s members mistake their equality for strength when it is chiefly an expensive way of ensuring that seventeen governments must be persuaded before anything can be done.

The thaw

Since the 1990s

The cold war did not end. It cooled, and the cooling has been slow, uneven and reversible, and the alliance has spent thirty years failing to decide what to do about it.

The 1990s produced the opening of the Russian archives, the first serious contacts between allied and Russian staffs, and the discovery of what the crisis of 1968 had actually been. That discovery is generally credited with the change in tone: the west learned that the closest it had come to a war was an accident produced by a Russian internal crisis, and Russia learned how close the automatic guarantee had brought sixteen states to acting on it.

A standing consultative arrangement between the alliance and Russia was established, has met continuously since, and has produced incident-prevention agreements, exercise notification, and a permanent channel at staff level that has never been suspended even in the worst years. It has produced nothing on the central question, because the central question is the guarantee, and the guarantee is not negotiable.

The unresolved present

Russia is in flux in a way that its neighbours find more unsettling than the confrontation was. In two days of October 2028 the authoritarian republic decreed a free election, elected a President, voted to restore the Romanovs and proclaimed an Emperor, and the alliance now faces a constitutional empire that has not yet written its constitution.

A great state whose direction is uncertain is harder to plan against than one whose intentions are hostile and known. The alliance’s members have divided, roughly along the same lines as everything else, between those who read the flux as an opportunity and those who read it as a warning. Nothing in the alliance’s structure helps: an organisation designed to be incapable of taking rapid decisions is also incapable of taking a view.

What the alliance has done, in the absence of agreement, is continue. The exercises run, the standardisation agreements accumulate, the ratification tables are published, the commander rotates, the Council meets, and Article IV sits where it has sat since 1949, widened once in its area and never in its working, untested, and automatic.

The rest of the treaty

The Treaty of Brussels runs to fourteen articles, of which two are famous and the remainder do the ordinary work of holding an alliance together.

Articles I to III are preambular in substance. They record that the parties act within the League Charter, that they will settle disputes among themselves peaceably, and that they will maintain and develop their individual and collective capacity to resist attack. The third of these is the only obligation in the treaty that a member can fail without an enemy being involved, and it is unenforceable: no member has ever been formally accused of failing to maintain its capacity, though several have been criticised in the Council for it.

Article IV is the guarantee. Article V defines the area to which it applies. As signed it was Europe, North America, and the vessels and aircraft of the parties upon the Atlantic; since Nigeria’s protocol of accession of 1999 it has been the territory of any of the parties, other than territories overseas, with their vessels and aircraft upon the Atlantic. The definition has been argued about continuously. It excludes the members’ overseas possessions and the waters of other oceans, which was deliberate: the founders would not commit the alliance to colonial wars, and the exclusion was insisted upon by the American delegation over the objections of two European members who had a great deal of overseas territory in 1949 and have very little now.

Article VI provides for consultation whenever a member considers its security threatened, and it is the article the alliance actually uses. Consultation carries no obligation beyond attendance, and it is invoked several times a year.

Article VII is the limitation, and is discussed above.

Articles VIII to XI are institutional: the Council, its power to establish subsidiary bodies, the Secretary-General, and the international staff. They are notably brief. The alliance’s institutional structure has been built almost entirely by Council decision rather than by treaty, which is why it has been possible to change it without ratification, and why the things that cannot be changed without ratification are exactly the things the founders wanted fixed.

Article XII governs accession: unanimous invitation, and ratification by each existing party.

Article XIII is the withdrawal clause. One year’s notice, in writing, to the depositary. It has never been used.

Article XIV provides that the English, French and German texts are equally authentic. The three texts have been found to diverge in four places, all of them minor, none of them in Articles IV or VII, which were drafted simultaneously in all three languages by a committee that met for eleven days and is the reason those two articles are the clearest prose in the instrument.

Chronology

Trivia

See also

• The League of Nations
• The European Concert
• The Russian Republic
• Harry S. Truman
• The May Day bombings
• The German Empire
• The Scandinavian Union
• The Federal Kingdom of Nigeria, the fourth enlargement
• The Royal American Armed Forces

Categories

World Institutions · Alliances · Defence · The Cold War · Treaties · Foreign Relations